CN114488190A - Laser radar 3D point cloud ground detection method - Google Patents
Laser radar 3D point cloud ground detection method Download PDFInfo
- Publication number
- CN114488190A CN114488190A CN202111650419.8A CN202111650419A CN114488190A CN 114488190 A CN114488190 A CN 114488190A CN 202111650419 A CN202111650419 A CN 202111650419A CN 114488190 A CN114488190 A CN 114488190A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- voxel grid
- ground
- laser radar
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
- G01S17/93—Lidar systems specially adapted for specific applications for anti-collision purposes
- G01S17/931—Lidar systems specially adapted for specific applications for anti-collision purposes of land vehicles
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
- G01S7/495—Counter-measures or counter-counter-measures using electronic or electro-optical means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Computer Networks & Wireless Communication (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Electromagnetism (AREA)
- Optical Radar Systems And Details Thereof (AREA)
Abstract
The invention discloses a laser radar 3D point cloud ground detection method. The problem that the expected ground filtering effect is difficult to achieve in the prior art is solved; the invention comprises the following steps: collecting point cloud data of roads around a vehicle; cutting point cloud data, and determining a region range of interest; initializing the 3D voxel grids, and obtaining the coordinates of each 3D voxel grid in a laser radar coordinate system; traversing all the point cloud data, filling the point cloud data into a 3D voxel grid corresponding to the coordinates, and recording indexes; performing plane fitting on the point cloud data in each 3D voxel grid, traversing all the point cloud data, calculating the distance from the point cloud data to a plane, comparing a comprehensive distance threshold value with a height threshold value, and judging whether the corresponding point cloud data is a ground point; and sending all point cloud data judged as non-ground points into a clustering algorithm to realize the detection of the target point of the obstacle. And ground points are identified and distinguished according to the scene of a single voxel, so that the expected ground filtering effect is achieved.
Description
Technical Field
The invention relates to the field of laser radar detection, in particular to a laser radar 3D point cloud ground detection method.
Background
In recent years, with the development of intelligent technology, the technology is applied to the fields of intelligent automobiles, robots and the like. Under certain environmental conditions, automatic driving of the vehicle can be realized. However, in order to realize the automatic driving technology, a plurality of sensors, such as a camera, a laser radar, a millimeter wave radar and the like, must be relied on to sense information around the vehicle, so as to control the vehicle to run in a safe range. What this patent relates to is lidar, and lidar can scan object 3D information, is an essential sensor to autopilot.
However, when the traditional target detection algorithm of the laser radar is applied, one of the key processes is the detection of ground points, and the quality of the ground detection result directly affects the target detection effect. Currently, a common ground detection algorithm is random sample consensus (RANSAC), and a plane fitting method is adopted to fit a plane containing the most ground points from a set of point cloud data. In practical application, the ground is not horizontal, and has a convex or concave situation, so that the expected ground filtering effect is difficult to achieve. Moreover, some algorithms are not applicable to solid state lidar, which can only use mechanical rotary lidar.
For example, a "road target 3D detection method and system based on lidar point cloud" disclosed in chinese patent literature has publication numbers: CN112883820A, published: 2021-06-01, comprising: receiving point cloud data acquired by a laser radar on a road; disassembling the point cloud data packet to obtain point cloud data of each frame, and respectively storing the point cloud data of each frame according to a time sequence; screening each frame of point cloud data according to the number threshold of the points in the selected target area so as to discard any frame of point cloud data with the number smaller than the number threshold; carrying out statistical filtering operation on the point cloud data; filtering the preprocessed point cloud data to determine target object point cloud data; clustering target object point cloud data to combine dense point cloud data to generate a point cloud target candidate set; and performing type judgment on the point cloud candidate set according to the pre-trained deep learning network model, and calculating the spatial position information of the target object.
The scheme is difficult to achieve the expected ground filtering effect, ground point cloud and non-ground point cloud data cannot be accurately classified, and the real noise reduction processing effect is achieved.
Disclosure of Invention
The invention mainly solves the problem that the prior art is difficult to achieve the expected ground filtering effect; the method comprises the steps of carrying out grid division on point cloud data through a 3D voxel grid, then carrying out operations such as plane fitting and the like to extract ground points, and achieving an expected ground filtering effect.
The technical problem of the invention is mainly solved by the following technical scheme:
a laser radar 3D point cloud ground detection method comprises the following steps:
s1: starting a vehicle-mounted laser radar, and collecting point cloud data of roads around a vehicle;
s2: point cloud data preprocessing; establishing a laser radar coordinate system, cutting point cloud data, and determining the range of the region of interest;
s3: initializing the 3D voxel grids, and obtaining the coordinates of each 3D voxel grid in a laser radar coordinate system;
s4: traversing all the point cloud data, filling the point cloud data into a 3D voxel grid corresponding to the coordinates, and recording indexes;
s5: performing plane fitting on the point cloud data in each 3D voxel grid, traversing all the point cloud data, calculating the distance from the point cloud data to a plane, comparing a comprehensive distance threshold value with a height threshold value, and judging whether the corresponding point cloud data is a ground point;
and S6, sending all point cloud data judged to be non-ground points into a clustering algorithm to realize the detection of the target point of the obstacle.
By adopting the scheme, the voxel division is carried out on the point cloud data, various complex scenes are converted into local simple scenes, the ground detection is carried out aiming at the simple scene information of a single voxel, the real situation of the ground can be better fitted, and the good ground filtering effect is achieved.
Preferably, the laser radar coordinate system takes the installation position of the laser radar as an origin, the positive front of the vehicle is the positive direction of an X axis, the left side of the vehicle is the positive direction of a Y axis, and the vertical upper part of the vehicle is the positive direction of a Z axis; the range x epsilon of the clipped region of interest (x1, x 2); y e (y1, y 2); z ∈ (z1, z 2). The range of the region of interest is adjusted through manual setting, the approximate scanning range of the laser radar is 100-200 meters, point cloud data possibly after 100 meters is very few and is not in the range of the algorithm, the range of the region of interest is adjusted through manual adaptability, good calculation data is obtained, remote noise point clouds are removed, and the operation efficiency of the algorithm is effectively improved.
Preferably, a single 3D voxel grid size is initialized to X0 × Y0 × Z0;
and obtaining the 3D voxel grid layout in the region of interest range according to the cut region of interest range and the preset single 3D voxel grid size:
GridSize v = (x2-x1) / X0 ；
GridSize u = (y2–y1) / Y0;
wherein GridSize v is the number of rows of the 3D voxel grid in the X-axis direction in the laser radar coordinate system;
GridSize u is the number of columns of the 3D voxel grid in the Y-axis direction in the lidar coordinate system.
Preferably, the method for determining the index value of the 3D voxel grid is as follows:
and sequentially sliding from left to right by taking the lower left corner of the 3D voxel grid as a starting position, sequentially sliding from left to right from the second row after the first row is finished until the first row slides to the upper right corner of the 3D voxel grid, and sequentially numbering to obtain the index value of the 3D voxel grid. Index values for the 3D voxel grid are obtained.
Preferably, the step S4 specifically includes the following steps:
calculating the offset from the laser radar coordinate system to the 3D voxel grid coordinate system;
obtaining the coordinate value of the converted point cloud data in the 3D voxel grid according to the offset;
dividing the coordinate values of the converted point cloud data by the size of the 3D voxel grid and rounding down to (u ', v');
obtaining an index value a of a 3D voxel grid where point cloud data is located according to the size of the 3D voxel grid:
a = u '× b + v', where b is the number of rows of the 3D voxel grid in which it is located;
and traversing all the point cloud data based on the laser radar coordinate system to obtain an index value a of the 3D voxel grid where each point cloud data is located. And obtaining an index value of a 3D voxel grid where the point cloud data is located through coordinate conversion.
Preferably, the step S5 specifically includes the following steps:
sorting the point cloud data in each 3D voxel grid according to the Z-axis coordinate value, and selecting the first 5 point cloud data of which the Z-axis coordinate value is smaller than a preset height threshold D1 and reaches the sorting from small to small;
performing plane fitting on the selected 5 point cloud data, and calculating the distance d from other point cloud data to a fitting plane;
if the value of the distance D is smaller than or equal to a preset distance threshold value D2, dividing the point cloud data into ground points, storing the index value of the point cloud data, and marking the mark bit as the ground point;
if the value of the distance D is larger than a preset distance threshold value D2, dividing the point cloud data into non-ground points, storing the index value of the point cloud data, and marking the mark bit as the non-ground point;
and traversing the whole 3D voxel grid, judging and marking all point cloud data as ground points or non-ground points, and respectively storing corresponding index values.
And identifying and distinguishing the ground points and the non-ground points to achieve the expected ground filtering effect.
Preferably, the step S5 specifically includes the following steps:
converting point cloud data in the 3D voxel grid into a two-dimensional coordinate system, and setting X values of all the point cloud data to be 0;
sorting the point cloud data from small to large according to Z-axis coordinate values, and selecting the first 5 point cloud data;
performing linear fitting on the Y coordinate value and the Z coordinate value of the selected 5 point cloud data;
calculating the distance from other point cloud data to the fitting straight line, if the calculated distance is smaller than a distance threshold D1, judging the point cloud data to be a ground point, otherwise, judging the point cloud data to be a non-ground point;
and traversing the whole 3D voxel grid, judging and marking all point cloud data as ground points or non-ground points, and respectively storing corresponding index values.
The ground points and the non-ground points are identified and distinguished, and the expected ground filtering effect is achieved.
The invention has the beneficial effects that:
1. by dividing the point cloud data into voxels, various complex scenes are converted into local simple scenes, and ground detection is performed on the simple scene information of a single voxel, so that the ground real situation can be better fitted, and the expected ground filtering effect is achieved.
2. The range of the region of interest is adjusted through cutting, good calculation data are obtained, remote noise point clouds are removed, and the operation efficiency of the algorithm is effectively improved.
Drawings
FIG. 1 is a flow chart of a laser radar 3D point cloud ground detection method of the present invention.
Fig. 2 is a schematic diagram of a 3D voxel grid of the present invention.
Detailed Description
The technical scheme of the invention is further specifically described by the following embodiments and the accompanying drawings.
The first embodiment is as follows:
the laser radar 3D point cloud ground detection method of the present embodiment, as shown in fig. 1, includes the following steps:
s1: and starting the vehicle-mounted laser radar, and collecting point cloud data of roads around the vehicle.
After the laser radar (mechanical or solid) is hung on a vehicle, the laser radar is started to drive, the laser radar acquires 3D point cloud data information of surrounding roads in the driving process of the vehicle, and the point cloud data is uploaded to a system and stored.
S2: point cloud data preprocessing; and establishing a laser radar coordinate system, cutting point cloud data, and determining the range of the region of interest.
After the received point cloud data is obtained, the point cloud data is preprocessed.
The preprocessing of the point cloud data includes cropping the point cloud data, leaving only the point cloud data within the region of interest. The range of the region of interest is adjusted according to the manual setting.
The point cloud data is based on a laser radar coordinate system which takes a laser radar installation position as an origin, the right front of the vehicle is the positive direction of an X axis, the left side of the vehicle is the positive direction of a Y axis, and the vertical upper part of the vehicle is the positive direction of a Z axis.
In the present embodiment, the range of interest for cropping is 0m < X <70 m; -30m < Y <30 m; -1m < Z <6 m.
S3: and initializing the 3D voxel grids, and obtaining the coordinates of each 3D voxel grid in a laser radar coordinate system.
As shown in fig. 2, the laser radar coordinate system is divided into several 3D voxel grids, and in the present embodiment, the size of each 3D voxel grid is 10 × 10 × 7 as an initialization value.
And obtaining the 3D voxel grid layout in the region of interest range according to the cut region of interest range and the preset 3D voxel grid size:
GridSize v = (70-0)/10 = 7 lines;
GridSize u = (30- (-30))/10 = 6 columns.
Performing index initialization on the obtained 3D voxel grids, and initializing an index value of each 3D voxel grid, wherein the index value initialization method comprises the following steps:
as shown in fig. 2, the 3D voxel grid sequentially slides from left to right with the lower left corner as the starting position, and after the first row is finished, the 3D voxel grid sequentially slides from left to right from the second row until the second row slides to the upper right corner of the 3D voxel grid, and in this embodiment, 42 index values with numbers 0 to 41 are finally obtained.
S4: and traversing all the point cloud data, filling the point cloud data into a 3D voxel grid corresponding to the coordinates, and recording the index.
The preprocessed point cloud data of the region of interest is obtained in step S2, and a 7 × 6 3D voxel grid is obtained in step S3. Traversing all point cloud data in the range of the region of interest, performing voxel division according to coordinates after coordinate transformation of all the point cloud data, filling the point cloud data into a corresponding 3D voxel grid, and recording indexes of the point cloud data and the index of the voxel where the point cloud data is located.
The method comprises the following specific steps:
the offset from the lidar coordinate system to the 3D voxel grid coordinate system is calculated.
Since the coordinates of the point cloud data are based on the lidar coordinate system (X, Y, Z) and the origin of the coordinate system (u, v) of the 3D voxel grid is at the lower left corner, the calculated offset amount in the X-axis direction is 0 and the calculated offset amount in the Y-axis direction is 30 in this embodiment.
Obtaining the coordinate of the converted point cloud data in the 3D voxel grid as (X, 30-Y) according to the offset;
dividing the coordinate values (X, 30-Y) of the converted point cloud data by the size of the 3D voxel grid and rounding down to (u ', v');
obtaining an index value a of a 3D voxel grid where point cloud data is located according to the size of the 3D voxel grid:
a = u '× b + v', where b is the number of rows of the 3D voxel grid in which it is located.
And traversing all the point cloud data based on the laser radar coordinate system to obtain an index value a of the 3D voxel grid where each point cloud data is located.
S5: and performing plane fitting on the point cloud data in each 3D voxel grid, traversing all the point cloud data, calculating the distance from the point cloud to the plane, comparing the comprehensive distance threshold with the height threshold, and judging whether the corresponding point cloud data is a ground point.
After the filling of the point cloud data of the 3D voxel grid in step S4 is completed, the index and coordinates of the point cloud data in each 3D voxel grid in the laser radar coordinate system can be known.
And sequencing the point cloud data in each 3D voxel grid according to the Z-axis coordinate value, and selecting the first 5 point cloud data with the Z-axis coordinate value smaller than a preset height threshold D1 and the smallest Z-axis coordinate value as ground points.
And performing plane fitting on the 5 point cloud data selected as the ground points, and calculating the distance d from other point cloud data to the fitting plane.
And judging whether the corresponding point cloud data is a ground point or not according to the value of the distance d. If the value of the distance D is smaller than or equal to a preset distance threshold value D2, dividing the point cloud data into ground points, storing the index value of the point cloud data, and marking the mark bit as the ground point; if the value of the distance D is larger than a preset distance threshold value D2, the point cloud data is divided into non-ground points, the index value of the point cloud data is stored, and the mark bit is marked as the non-ground point.
And traversing the whole 3D voxel grid, judging and marking all point cloud data as ground points or non-ground points, and respectively storing corresponding index values.
And S6, sending all point cloud data judged to be non-ground points into a clustering algorithm to realize the detection of the target point of the obstacle.
Extracting ground points, obtaining index values which are judged as the ground points in all the point cloud data through the step S5, and obtaining all the ground points according to the index values; ground points are removed through index values, and European clustering operation is carried out on non-ground points, so that the detection of the target point of the obstacle can be realized.
According to the scheme of the embodiment, the original point cloud data is subjected to voxel grid division, so that an overall complex scene is converted into a local ideal scene. Through multi-condition threshold and plane fitting, classification of ground points and non-ground points is effectively achieved.
Example two:
the scheme of the present embodiment optimizes step S5.
After the index and the coordinates of the point cloud data in each 3D voxel grid under the laser radar coordinate system are obtained through the process of step S4, the point cloud data in the 3D voxel grid is converted into points on the two-dimensional coordinate system, that is, all X values are set to 0.
Sorting the point cloud data from small to large according to Z-axis coordinate values, and selecting the first five point cloud data;
performing straight line fitting on Y, Z coordinate values of the selected five point cloud data, and if there are ground points, fitting a straight line;
and calculating the distance from other point cloud data to the fitting straight line, if the calculated distance is less than a distance threshold D1, judging that the point cloud data is a ground point, otherwise, judging that the point cloud data is a non-ground point.
The scheme of this embodiment optimizes step S5, and the rest is the same as in the first embodiment.
It should be understood that the examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Further, it should be understood that various changes or modifications of the present invention may be made by those skilled in the art after reading the teaching of the present invention, and such equivalents may fall within the scope of the present invention as defined in the appended claims.
Claims (7)
1. A laser radar 3D point cloud ground detection method is characterized by comprising the following steps:
s1: starting a vehicle-mounted laser radar, and collecting point cloud data of roads around a vehicle;
s2: point cloud data preprocessing; establishing a laser radar coordinate system, cutting point cloud data, and determining the range of the region of interest;
s3: initializing the 3D voxel grids, and obtaining the coordinates of each 3D voxel grid in a laser radar coordinate system;
s4: traversing all the point cloud data, filling the point cloud data into a 3D voxel grid corresponding to the coordinates, and recording indexes;
s5: performing plane fitting on the point cloud data in each 3D voxel grid, traversing all the point cloud data, calculating the distance from the point cloud data to a plane, comparing a comprehensive distance threshold value with a height threshold value, and judging whether the corresponding point cloud data is a ground point;
and S6, sending all point cloud data judged to be non-ground points into a clustering algorithm to realize the detection of the target point of the obstacle.
2. The method of claim 1, wherein the lidar coordinate system is based on a lidar mounting position, the positive direction of an X axis is directly in front of the vehicle, the positive direction of a Y axis is directly on the left side of the vehicle, and the positive direction of a Z axis is vertically above the vehicle; the range x epsilon of the clipped region of interest (x1, x 2); y e (y1, y 2); z ∈ (z1, z 2).
3. The laser radar 3D point cloud ground detection method according to claim 1 or 2, wherein a single 3D voxel grid size is initialized to be X0 xY 0 xZ 0;
obtaining the 3D voxel grid layout in the region of interest range according to the cut region of interest range and the preset size of a single 3D voxel grid:
GridSize v = (x2-x1) / X0 ；
GridSize u = (y2–y1) / Y0;
wherein GridSize v is the number of rows of the 3D voxel grid in the X-axis direction in the laser radar coordinate system;
GridSize u is the number of columns of the 3D voxel grid in the Y-axis direction in the lidar coordinate system.
4. The ground detection method for the laser radar 3D point cloud according to claim 3, wherein the method for determining the index value of the 3D voxel grid comprises the following steps:
and sequentially sliding from left to right by taking the lower left corner of the 3D voxel grid as a starting position, sequentially sliding from left to right from the second row after the first row is finished until the first row slides to the upper right corner of the 3D voxel grid, and sequentially numbering to obtain the index value of the 3D voxel grid.
5. The method for ground detection by laser radar 3D point cloud according to claim 1, wherein the step S4 specifically includes the following steps:
calculating the offset from the laser radar coordinate system to the 3D voxel grid coordinate system;
obtaining the coordinate value of the converted point cloud data in the 3D voxel grid according to the offset;
dividing the coordinate value of the converted point cloud data by the size of the 3D voxel grid and rounding down to (u ', v');
obtaining an index value a of a 3D voxel grid where point cloud data is located according to the size of the 3D voxel grid:
a = u '× b + v', where b is the number of rows of the 3D voxel grid in which it is located;
and traversing all the point cloud data based on the laser radar coordinate system to obtain an index value a of the 3D voxel grid where each point cloud data is located.
6. The method for detecting the ground by the laser radar 3D point cloud according to claim 1, 4 or 5, wherein the step S5 specifically includes the following steps:
sorting the point cloud data in each 3D voxel grid according to the Z-axis coordinate value, and selecting the first 5 point cloud data of which the Z-axis coordinate value is smaller than a preset height threshold D1 and reaches the sorting from small to small;
performing plane fitting on the selected 5 point cloud data, and calculating the distance d from other point cloud data to a fitting plane;
if the value of the distance D is smaller than or equal to a preset distance threshold value D2, dividing the point cloud data into ground points, storing the index value of the point cloud data, and marking the mark bit as the ground point;
if the value of the distance D is larger than a preset distance threshold value D2, dividing the point cloud data into non-ground points, storing the index value of the point cloud data, and marking the mark bit as the non-ground point;
and traversing the whole 3D voxel grid, judging and marking all point cloud data as ground points or non-ground points, and respectively storing corresponding index values.
7. The method for detecting the ground by the laser radar 3D point cloud according to claim 1, 4 or 5, wherein the step S5 specifically includes the following steps:
converting point cloud data in the 3D voxel grid into a two-dimensional coordinate system, and setting X values of all the point cloud data to be 0;
sorting the point cloud data from small to large according to Z-axis coordinate values, and selecting the first 5 point cloud data;
performing linear fitting on the Y coordinate value and the Z coordinate value of the selected 5 point cloud data;
calculating the distance from other point cloud data to the fitting straight line, if the calculated distance is smaller than a distance threshold D1, judging the point cloud data to be a ground point, otherwise, judging the point cloud data to be a non-ground point;
and traversing the whole 3D voxel grid, judging and marking all point cloud data as ground points or non-ground points, and respectively storing corresponding index values.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111650419.8A CN114488190A (en) | 2021-12-30 | 2021-12-30 | Laser radar 3D point cloud ground detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111650419.8A CN114488190A (en) | 2021-12-30 | 2021-12-30 | Laser radar 3D point cloud ground detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114488190A true CN114488190A (en) | 2022-05-13 |
Family
ID=81507581
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111650419.8A Pending CN114488190A (en) | 2021-12-30 | 2021-12-30 | Laser radar 3D point cloud ground detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114488190A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114820662A (en) * | 2022-05-23 | 2022-07-29 | 燕山大学 | Road side visual angle ground segmentation method, system and medium based on point cloud two-dimensional density |
| CN116071571A (en) * | 2023-03-03 | 2023-05-05 | 北京理工大学深圳汽车研究院(电动车辆国家工程实验室深圳研究院) | Robust and rapid vehicle single-line laser radar point cloud clustering method |
| CN116229405A (en) * | 2023-05-05 | 2023-06-06 | 倍基智能科技(四川)有限公司 | Method for detecting ground from point cloud data |
| CN116413740A (en) * | 2023-06-09 | 2023-07-11 | 广汽埃安新能源汽车股份有限公司 | Laser radar point cloud ground detection method and device |
| CN116627164A (en) * | 2023-04-13 | 2023-08-22 | 北京数字绿土科技股份有限公司 | Terrain-height-based unmanned aerial vehicle ground-simulated flight control method and system |
-
2021
- 2021-12-30 CN CN202111650419.8A patent/CN114488190A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114820662A (en) * | 2022-05-23 | 2022-07-29 | 燕山大学 | Road side visual angle ground segmentation method, system and medium based on point cloud two-dimensional density |
| CN116071571A (en) * | 2023-03-03 | 2023-05-05 | 北京理工大学深圳汽车研究院(电动车辆国家工程实验室深圳研究院) | Robust and rapid vehicle single-line laser radar point cloud clustering method |
| CN116627164A (en) * | 2023-04-13 | 2023-08-22 | 北京数字绿土科技股份有限公司 | Terrain-height-based unmanned aerial vehicle ground-simulated flight control method and system |
| CN116627164B (en) * | 2023-04-13 | 2024-04-26 | 北京数字绿土科技股份有限公司 | Terrain-height-based unmanned aerial vehicle ground-simulated flight control method and system |
| CN116229405A (en) * | 2023-05-05 | 2023-06-06 | 倍基智能科技(四川)有限公司 | Method for detecting ground from point cloud data |
| CN116413740A (en) * | 2023-06-09 | 2023-07-11 | 广汽埃安新能源汽车股份有限公司 | Laser radar point cloud ground detection method and device |
| CN116413740B (en) * | 2023-06-09 | 2023-09-05 | 广汽埃安新能源汽车股份有限公司 | Laser radar point cloud ground detection method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108932736B (en) | Two-dimensional laser radar point cloud data processing method and dynamic robot pose calibration method | |
| CN114488190A (en) | Laser radar 3D point cloud ground detection method | |
| CN110320504B (en) | Unstructured road detection method based on laser radar point cloud statistical geometric model | |
| CN108828621A (en) | Obstacle detection and road surface partitioning algorithm based on three-dimensional laser radar | |
| CN106650640B (en) | Negative obstacle detection method based on laser radar point cloud local structure characteristics | |
| CN111583337A (en) | Omnibearing obstacle detection method based on multi-sensor fusion | |
| CN112785643A (en) | Indoor wall corner two-dimensional semantic map construction method based on robot platform | |
| CN108460416A (en) | A kind of structured road feasible zone extracting method based on three-dimensional laser radar | |
| CN112396650A (en) | Target ranging system and method based on fusion of image and laser radar | |
| CN109738910A (en) | A kind of curb detection method based on three-dimensional laser radar | |
| CN115049700A (en) | Target detection method and device | |
| CN112487919A (en) | 3D target detection and tracking method based on camera and laser radar | |
| CN104880160B (en) | Two-dimensional-laser real-time detection method of workpiece surface profile | |
| CN116229408A (en) | Target identification method for fusing image information and laser radar point cloud information | |
| CN114972968A (en) | Tray identification and pose estimation method based on multiple neural networks | |
| CN116109601A (en) | Real-time target detection method based on three-dimensional laser radar point cloud | |
| Zelener et al. | Cnn-based object segmentation in urban lidar with missing points | |
| CN115797813B (en) | Water environment pollution detection method based on aerial image | |
| CN115620261A (en) | Vehicle environment sensing method, system, equipment and medium based on multiple sensors | |
| CN116486287A (en) | Target detection method and system based on environment self-adaptive robot vision system | |
| CN114463362A (en) | Three-dimensional collision avoidance sonar obstacle detection method and system based on deep learning | |
| CN114821526A (en) | Obstacle three-dimensional frame detection method based on 4D millimeter wave radar point cloud | |
| CN113820682B (en) | Millimeter wave radar-based target detection method and device | |
| CN115327572A (en) | Method for detecting obstacle in front of vehicle | |
| CN111880197A (en) | Road boundary detection method based on four-line laser radar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |